Chuck Norcutt wrote:
>Since noise is worse with low signal to noise ratio how is noise reduced
>by cutting the signal strength to each sensor by a factor of at least 3?
>
>
I think the whole idea is a nonstarter for all kinds of reasons. Not the
least is the cost of 3 sensors and the size, weight, and mechanical
complexity of such a setup. Certainly not practical for any consumer
size/price camera. It's essentially what the Foveon sensor does
electronically, but done mechanically.
However, I'm not sure the issue you raise is a problem. With an RGB
sensor, about 2/3 of the light that reaches each individual sensor site
is filtered out by the individual sensor site filter. With the
beam/color splitter model, the light is separated en masse, rather than
at the sensor level.
Assuming roughly equal general losses in the two different filtering
strategies, the separate sensor model might even be better, as all the
light at each xy coordinate would be sensed, rather than 2/3 being
discarded. How it would all work out compared to a Bayer process, I have
no idea, but it is not inherently working with lower light levels
actually being sensed at each site and does have more total light being
processed.
Moose
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|